You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Lots of work to make this happen. Adding a non-unique column should be a straightforward change to the TableDescriptor. Adding a unique column would also require adding a unique index which will require the infrastructure for the multi-step process of:
Adding the index to the TableDescriptor but marking it as write-only.
Waiting for the TableDescriptor change to propagate to all of the nodes.
Backfilling the index with some sort of mapreduce over the table.
Marking the index as read-write in the TableDescriptor.
Dropping a column (assuming it is not referenced by an index) will also be a multi-step process:
Marking the column as deleting in the TableDescriptor.
Waiting for the TableDescriptor change to propagate to all of the nodes.
Deleting the column with some sort of mapreduce over the table.
Removing the column from the TableDescriptor.
The text was updated successfully, but these errors were encountered:
Lots of work to make this happen. Adding a non-unique column should be a straightforward change to the
TableDescriptor
. Adding a unique column would also require adding a unique index which will require the infrastructure for the multi-step process of:TableDescriptor
but marking it as write-only.TableDescriptor
change to propagate to all of the nodes.TableDescriptor
.Dropping a column (assuming it is not referenced by an index) will also be a multi-step process:
TableDescriptor
.TableDescriptor
change to propagate to all of the nodes.TableDescriptor
.The text was updated successfully, but these errors were encountered: